Enable AI to control your browser

Deutsch | Español | français | 日本語 | 한국어 | Português | Русский | 中文

🤖 Quickstart

With uv (Python>=3.11):

#  We ship every day - use the latest version!
uv pip install browser-use

Download chromium using playwright's shortcut:

uvx playwright install chromium --with-deps --no-shell

Create a .env file and add your API key. Don't have one? Start with a free Gemini key.

GEMINI_API_KEY=

Run your first agent:

from browser_use import Agent, ChatGoogle
from dotenv import load_dotenv
load_dotenv()

agent = Agent(
    task="Find the number of stars of the browser-use repo",
    llm=ChatGoogle(model="gemini-flash-latest"),
    # browser=Browser(use_cloud=True),  # Uses Browser-Use cloud for the browser
)
agent.run_sync()

Check out the library docs and cloud docs for more settings.

Stealth Browser Infrastructure

Want to bypass Cloudflare, or any other anti-bot protection?

Simply go to Browser Use Cloud grab a BROWSER_USE_API_KEY and use the use_cloud parameter.

from browser_use import Agent, Browser, ChatOpenAI

# Use Browser-Use cloud browser service
browser = Browser(
    use_cloud=True,  # Automatically provisions a cloud browser
)

agent = Agent(
    task="Your task here",
    llm=ChatOpenAI(model='gpt-4.1-mini'),
    browser=browser,
)

Demos

Task: Add grocery items to cart, and checkout.

Task: Read my CV & find ML jobs, save them to a file, and then start applying for them in new tabs, if you need help, ask me.

apply.to.jobs.8x.mp4

See more examples and give us a star!

MCP Integration

This gives Claude Desktop access to browser automation tools for web scraping, form filling, and more. See the MCP docs.

{
  "mcpServers": {
    "browser-use": {
      "command": "uvx",
      "args": ["browser-use[cli]", "--mcp"],
      "env": {
        "OPENAI_API_KEY": "sk-..."
      }
    }
  }
}

Tell your computer what to do, and it gets it done.

Made with ❤️ in Zurich and San Francisco